INDEX

Explanations

designation -er words

The neuron is sensitive to morphologically complex, affix-bearing content words—especially hyphenated compounds and words built with prefixes or suffixes (e.g. “cookie-cutter,” “dimpled,” “self-titled,” “embedded,” “sed,” “pedometer,” “zeros”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

</h3>

-3.17

if

-1.84

꜆

-1.82

</sub>

-1.77

」「

-1.73

must

-1.73

沨

-1.63

³

-1.59

 ที่

-1.58

 たん

-1.58

POSITIVE LOGITS

2.22

and

1.52

</b>

1.39

IZATION

1.38

sandalia

1.37

URE

1.35

 trabaj

1.34

 koš

1.34

apnews

1.33

飪

1.33

Activations Density 0.042%